Skip to content

Latest commit

 

History

History
59 lines (40 loc) · 2.32 KB

File metadata and controls

59 lines (40 loc) · 2.32 KB

TensorFlow Lite Image Classification Demo

Overview

This is a camera app that continuously classifies the objects in the frames seen by your device's back camera, with the option to use a quantized MobileNet V1, EfficientNet Lite0, EfficientNet Lite1, or EfficientNet Lite2 model trained on Imagenet (ILSVRC-2012-CLS). These instructions walk you through building and running the demo on an Android device.

The model files are downloaded via Gradle scripts when you build and run the app. You don't need to do any steps to download TFLite models into the project explicitly.

This application should be run on a physical Android device.

App example showing UI controls. Result is espresso.

App example without UI controls. Result is espresso.

This sample demonstrates how to use TensorFlow Lite with Kotlin. If you would like to see an example using Java, please go to the android_java sample directory.

Build the demo using Android Studio

Prerequisites

  • The Android Studio IDE (Android Studio 2021.2.1 or newer). This sample has been tested on Android Studio Chipmunk

  • A physical Android device with a minimum OS version of SDK 23 (Android 6.0 - Marshmallow) with developer mode enabled. The process of enabling developer mode may vary by device.

Building

  • Open Android Studio. From the Welcome screen, select Open an existing Android Studio project.

  • From the Open File or Project window that appears, navigate to and select the tensorflow-lite/examples/image_classification/android directory. Click OK.

  • If it asks you to do a Gradle Sync, click OK.

  • With your Android device connected to your computer and developer mode enabled, click on the green Run arrow in Android Studio.

Models used

Downloading, extraction, and placing the models into the assets folder is managed automatically by the download.gradle file.